K-Means Clustering Visualization of Web-Based OLAP Operations for Hotspot Data

Abstract

Jn the pre~lous work we dtl·cloped the web-based OLAP (On-Une Analytical P~lng) integrated 141tb the data 14•arehouse tor hotspot data in Indonesia. This work aims to develop 11 vlsullllzatlon module for houpot clusten ~ulted from OLAP operations including roll up and drill do14n. The daui warehouse consists or hotspot da1:1 represented ln multidimensional model with two dimensions: time and k>cation. In the dimension time, the ordered sequence or elements rrom the higher-level of hierarchy to tJ1e lowest Is from year, quarter, to month. Wbtrtas, the M>quence In the dimension location Is from Wand, province, to district. The clustering algorithm we n1lplled was K-means In which l11e bcsl clustering 14IL~ obtoined for the size or cluster 4 with 8\-CTagl' \-aluc or SSE (sum or square error) 0.2944 ror combinations of clements in the dimension time and locado11. llotspot clusters arc 'isualizcd In form of maps In addlUoo to crosstam and graphics buill in the pre,lous 140rk. The map module in the web-based OLAP can be used to belier org1111.lie and analyze the hotspot data as one or indicators for forest f1ru occurrence In Indonesia.